Research map / Image generation / Text-to-image diffusion
DiverseGRPO: Mitigating Mode Collapse in Image Generation via Diversity-Aware GRPO
Henglin Liu, Huijuan Huang, Jing Wang, Chang Liu, Xiu Li, Xiangyang Ji
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add distributional creativity bonus and structure-aware regularization to Flow-GRPO | SD3.5-M / PickScore · SD3.5-M / PickScore reward | DreamSim: 0.1278 -> 0.1517 | Improved |
| add DiverseGRPO to Flow-GRPO | SD3.5-M / PickScore · SD3.5-M / PickScore reward | FID: 56.206 -> 43.115 | Improved |
| add DiverseGRPO to Flow-GRPO | SD3.5-M / PickScore · SD3.5-M / PickScore reward | BeyondFID: 0.0667 -> 0.1895 | Improved |
| add DiverseGRPO to Flow-GRPO | SD3.5-M / PickScore · SD3.5-M / PickScore reward | ImageReward: 1.3650 -> 1.4009 | Improved |
| add DiverseGRPO to Flow-GRPO | Flux.1-dev / PickScore · Flux.1-dev / PickScore reward | DreamSim: 0.1382 -> 0.1575 | Improved |
| add DiverseGRPO to Flow-GRPO | Flux.1-dev / PickScore · Flux.1-dev / PickScore reward | FID: 68.746 -> 62.505 | Improved |
| add DiverseGRPO to Flow-GRPO | Flux.1-dev / PickScore · Flux.1-dev / PickScore reward | PickScore: 0.8750 -> 0.8779 | No clear effect |
| add DiverseGRPO to Flow-GRPO | SD3.5-M / HPSv3 · SD3.5-M / HPSv3 reward | DreamSim: 0.1625 -> 0.1851 | Improved |
| add DiverseGRPO to Flow-GRPO | SD3.5-M / HPSv3 · SD3.5-M / HPSv3 reward | BeyondFID: 0.0971 -> 0.1646 | Improved |
| add DiverseGRPO to Flow-GRPO | SD3.5-M / HPSv3 · SD3.5-M / HPSv3 reward | PickScore: 0.8445 -> 0.8462 | No clear effect |
| add structure-aware regularization alone to GRPO (ablation) | DiverseGRPO · SD3.5-M | diversity | Improved |
| add creativity reward on top of SA-Reg (ablation) | DiverseGRPO · SD3.5-M | diversity | Improved |
Closest papers by meaning
- Beyond the Dirac Delta: Mitigating Diversity Collapse in Reinforcement Fine-Tuning for Image Generation (NeurIPS 2026 · similarity 0.89)
- Training Diffusion Models Towards Diverse Image Generation with Reinforcement Learning (CVPR 2024 · similarity 0.89)
- Reinforcement Learning Meets Masked Generative Models: Mask-GRPO for Text-to-Image Generation (NeurIPS 2025 · similarity 0.89)
- Delving into RL for Image Generation with CoT: A Study on DPO vs. GRPO (NeurIPS 2025 · similarity 0.88)
- MaskSense: Confronting the Visual Exploration Trap in Masked Image Generation (NeurIPS 2026 · similarity 0.88)
- Co-GRPO: Co-Optimized Group Relative Policy Optimization for Masked Image Generation (NeurIPS 2026 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
